The Local and Systemic Exposure to Oxygen in Children With Severe Bronchiolitis on Invasive Mechanical Ventilation: A Retrospective Cohort Study : Pediatric Critical Care Medicine

Journal Logo

Online Brief Report

The Local and Systemic Exposure to Oxygen in Children With Severe Bronchiolitis on Invasive Mechanical Ventilation: A Retrospective Cohort Study

Lilien, Thijs A. MD; de Sonnaville, Eleonore S. V. MD; van Woensel, Job B. M. MD, PhD; Bem, Reinout A. MD, PhD

Author Information
Pediatric Critical Care Medicine 24(2):p e115-e120, February 2023. | DOI: 10.1097/PCC.0000000000003130


Viral-induced bronchiolitis is among the most important injuries to the pulmonary system in early life (1,2). Although generally a mild disease, children admitted to the PICU for severe bronchiolitis frequently require supplemental oxygen with invasive mechanical ventilation (IMV) to prevent and resolve life-threatening hypoxemia. However, toxic effects of prolonged high-dose oxygen exposure in adults and neonates have been well established (3,4). Recently, a meta-analysis of studies in critically ill children found an association between a high systemic oxygen burden, as reflected by arterial hyperoxemia, and mortality (5). This further underlines the potential risk of oxygen treatment beyond the direct postnatal period. Consequently, in light of the pro-inflammatory pulmonary microenvironment associated with severe bronchiolitis (6), high-dose oxygen therapy may predispose these patients to a second injury.

The current practice of oxygen titration in children with (severe) bronchiolitis is still largely based on expert opinion, while international guidelines vary in their recommendation for oxygenation targets (7,8). Better insight in the exposure to oxygen is needed to develop future study protocols on the potential role of oxygen toxicity in this vulnerable population in the PICU. In the current study, we aimed to assess the overall pulmonary (local) and arterial (systemic) exposure to oxygen in children with severe bronchiolitis receiving IMV. Our secondary aim was to estimate potentially avoidable exposure to high-dose oxygen.


This study was exempted by the Institutional Review Board of the Amsterdam UMC, location University of Amsterdam W22_127#22.168 (April 7, 2022).

Patients less than 2 years old who were admitted to the PICU, between December 2008 and December 2020, for severe bronchiolitis and required IMV were included.

Patient records were retrieved based at admission codes (eTable 1, and reviewed for eligibility. In brief, extracted data included: patient characteristics; outcome data; hourly nurse-validated ventilator data up to the 10th cumulative day of IMV; and results from arterial blood gas analyses (eMethods, Patients who were on IMV less than 24 hours, or more than 24 hours before admission, or those temporarily transferred to another PICU were excluded due to incomplete available data.

Fio2 and Pao2 were used to assess local and systemic oxygen exposure, respectively. Fio2 greater than or equal to 0.50 was defined as high, Fio2 0.30–0.49 as moderate, and Fio2 0.21–0.29 as low exposure (9). Pao2 greater than 100 Torr (> 13.3 kPa) was defined as supraphysiological and greater than 248 Torr (> 33 kPa) as severe hyperoxemia (5). To quantify possible excessive oxygen exposure, the cumulative excessive oxygen exposure score (CEE) was calculated (10). Oxygen administered above room air when peripheral oxygen saturation (Spo2) was greater than or equal to 97% was scored as excessive (11,12). A sensitivity analysis was included with Spo2 greater than or equal to 95% (eMethods, (10). CEE of 0.17–0.25 was defined as high and greater than 0.25 as severe exposure (10). We also assessed the healthcare worker’s response on Fio2 settings if Spo2 was greater than or equal to 97% by assessing any change in Fio2 during the next hour of IMV.

For a comprehensive description of the statistical analysis, see eMethods ( In summary, time-weighted averages (TWAs) were calculated per variable of interest for the duration of IMV (TWAIMV) and consecutive days of IMV (24-hr window [TWA24h]). Cases with substantial missing data were omitted. Ventilator-related data were assessed using a mixed-effects model. Fio2 and CEE score were evaluated using repeated measures correlation (13). Finally, we performed exploratory analyses on arterial line use and if TWA24h–Fio2 and TWA24h–CEE on the first day of IMV were associated with duration of IMV and length of stay (LOS) in the PICU. Analyses were performed using R 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria) with RStudio 1.3.1093 (RStudio, Boston, MA).


The selection of records is shown in eFigure 1 ( A total of 176 patients were included in the analysis. Descriptive statistics are reported in eTable 2 (

A total of 24,451 hourly time points of IMV were observed, generating 987 TWA24h’s. A negligible proportion of observed time points lacked data (Spo2: n = 672, 2.7%; Fio2: n = 868, 3.5%). In total, 2,007 single Pao2 values could be extracted (546 TWA24h’s). Exploratory analysis showed that the use of arterial lines in our center depended at admission year (p < 0.001) and decreased over time (eFig. 2,

Overall pulmonary oxygen exposure was moderate (median TWAIMV–Fio2 0.41 [interquartile range (IQR) 0.36–0.47]) and overall systemic exposure was within physiologic limits (median TWAIMV–Pao2 75 Torr [10.0 kPa] [IQR, 69–84]). In more detail, pulmonary exposure was highest on the first day of IMV (median TWA24h–Fio2 0.46 [IQR, 0.39–0.53]) and significantly decreased during IMV (Fig. 1). However, short-term high-dose oxygen therapy was common among patients, with 89 patients (50.6%) having more than or equal to 1 day of high exposure, of which 25 patients (14.2%) had more than 3 days of high exposure (Table 1). Yet, none of these patients were exposed to severe hyperoxemia.

TABLE 1. - Summary of Time-Weighted Averages per 24-Hour Window of Ventilator-Related Data
Variable (Time-Weighted Averages per 24-hr Window) n Days a n Subjects b , 1 d n Subjects b , 2–3 d n Subjects b , > 3 d Total Subjects b , ≥ 1 d
Spo 2, n (%)
 < 92% 9 (0.9) 3 (1.7) 2 (1.1) 0 5 (2.8)
Fio 2, n (%)
 0.3–0.49 695 (70.4) 12 (6.8) 65 (36.9) 95 (54.0) 172 (97.7)
 ≥ 0.50 246 (24.9) 32 (18.2) 32 (18.2) 25 (14.2) 89 (50.6)
Pao 2 c , n (%)
 > 100 Torr (13.3 kPa) 38 (7.0) d 22 (15.7) e 7 (5.0) e 0 29 (20.7) e
 > 248 Torr (33 kPa) 0 NA NA NA 0
CEE (Spo 2 ≥ 97%) f , n (%)
 0.17–0.25 131 (13.3) 63 (35.8) 26 (14.8) 3 (1.7) 92 (52.3)
 > 0.25 46 (4.7) 15 (8.5) 8 (4.5) 3 (1.7) 26 (14.8)
CEE (Spo 2 ≥ 95%) g , n (%)
 0.17–0.25 283 (28.7) 52 (29.5) 68 (38.6) 16 (9.1) 136 (77.3)
 > 0.25 174 (17.6) 41 (23.3) 27 (15.3) 15 (8.5) 83 (47.2)
CEE = cumulative excessive oxygen exposure score, NA = not applicable, Spo2 = peripheral oxygen saturation.
aTotal observed days was 987.
bTotal patients was 176.
cData on Pao2 was only extracted from patients with an intra-arterial line.
dTotal observed days was 546.
eTotal patients with an intra-arterial line was 140.
fPrimary threshold used to define overuse of oxygen.
gThreshold of sensitivity analysis to define overuse of oxygen.
Number of days and number of subjects that fulfill the different criteria used to categorize oxygenation, pulmonary (local), arterial (systemic), and excessive oxygen exposure, respectively.

Figure 1.:
Time-weighted average per consecutive day (24 hr) (TWA24h) of invasive mechanical ventilation (IMV), TWAs were calculated by dividing the area under the variable by the observed time period. A, Fio 2, B, Pao 2, C, peripheral oxygen saturation (Spo 2), D, cumulative excessive oxygen exposure score (CEE) (calculated by the Fio 2 that was administered above room air [0.21], when oxygenation was sufficient, i.e., Spo 2 ≥ 97%). Dashed and black bars represent the median and interquartile range, respectively. The area of the violin plot represents the density of individual data points scaled to the number of patients (No.) (displayed below violin plot). *p < 0.05 versus day 1 of IMV, calculated using a mixed-effects model.

Potential overuse of supplemental oxygen was frequently observed: 45.9% (n = 3,118) of time points with high-oxygen exposure were seen at Spo2 greater than or equal to 97%. Quantification of overall possible excess oxygen showed a median TWAIMV–CEE of 0.11 (IQR, 0.09–0.14). Possible excessive use was also highest on the first day of IMV (median TWA24h–CEE 0.14 [IQR, 0.11–0.19]) and significantly decreased during IMV (Fig. 1). Short-term high and severe possible excessive oxygen exposure were commonly observed (Table 1). Sensitivity analysis, using a lowered Spo2 threshold, substantially increased the proportion of patients exposed to possible excessive oxygen use (Table 1). Furthermore, use of higher oxygen doses correlated moderately to highly with increasing excessive oxygen exposure (eFig. 3, Additionally, in 50.6% (n = 1,531) of the observed time points that patients were exposed to high-dose oxygen and Spo2 was greater than or equal to 97%, the Fio2 was left unchanged during the next hour of IMV (eFig. 4,

Last, exploratory analysis showed an increase of TWA24h–d1–Fio2 was associated with an increase in median days on IMV and LOS in the PICU. We failed to observe this association for TWA24h–d1–CEE, except in sensitivity analysis, where TWA24h–d1–CEE95% showed an association with an increase in median days LOS in the PICU (eTable 3,


In this retrospective cohort of 176 children with severe bronchiolitis who received IMV, we found that the local exposure to oxygen was moderate to high in the majority of patients, whereas the systemic burden of oxygen was relatively low. Furthermore, in many cases, oxygen exposure was deemed potentially excessive and probably avoidable. We found that increasing oxygen dosage correlated with an increase in potential oxygen overuse.

The observed pulmonary oxygen exposure resembled the liberal oxygenation arm of the Oxygen in Pediatric Intensive Care (Oxy-PICU) pilot trial (14). Yet, this level of pulmonary exposure did not result in a high systemic burden, which is in line with studies in adult acute hypoxemic respiratory failure (15). This discrepancy in exposure is likely attributable to intrapulmonary shunting, as observed in acute respiratory distress syndrome, a common complication of severe bronchiolitis (16). This highlights that we should not focus solely on Pao2 to assess oxygen toxicity (5), as this marker fails to address potential adverse effects of high pulmonary exposure in patients with injured lungs (4,9,10,17).

The observed tendency to treat patients with high-oxygen dosages was also observed in prior observational studies in critically ill adults (18,19). Although avoiding overzealous oxygen exposure is an integral part of protective ventilation strategies, healthcare workers apparently experience difficulties following such recommendations (18,19). Small-scale studies have demonstrated potential benefit of automated Fio2 closed-loop ventilation systems to improve Spo2 target adherence and reduce hyperoxia exposure (20,21).

High-dose pulmonary exposure and overuse of oxygen have previously been associated with worse outcome independent of disease severity (10,17,22). This partially aligns with our findings, as initial Fio2 levels were associated with prolonged disease, but for possible overuse, we observed an association with outcome only in sensitivity analysis. However, it is important to emphasize that our study was not designed to analyze outcomes. Regarding the potential benefit of reducing oxygen exposure, a multitude of trials in critically ill adults and preterm neonates have not demonstrated a clear benefit of restrictive oxygenation targets so far (15,23,24). Furthermore, overly restrictive targets have even been identified as potentially dangerous in preterm neonates (24). The ongoing Oxy-PICU trial will provide first evidence on the effects, including on long-term outcomes, in children (25).

Strengths of this study were the comprehensive hourly data collection and exposure analysis. Limitations include: first, its retrospective design, which is inherently associated with selection bias and unaccounted confounders. This may have particularly affected the observed systemic exposure. Second, local protocols for oxygen titration may differ between centers, hindering the generalizability of our results. Third, albeit based on previous reports (10,12,18), in the absence of optimal Spo2 targets, our definition of excessive oxygen administration remains arbitrary. Fourth, this study was not primarily designed to analyze associations with outcome and our observations are therefore only of an exploratory nature.


Moderate to high-dose pulmonary oxygen exposure and potential overuse of oxygen are common in children with severe bronchiolitis on IMV. As both of these have previously been associated with worse outcome in critically ill children, future studies on optimal oxygenation targets to reduce oxygen exposure are needed. In this regard, our findings help to develop study protocols for future trials in this field.


1. Agusti A, Faner R: Lung function trajectories in health and disease. Lancet Respir Med 2019; 7:358–364
2. Bem RA, Bont LJ, van Woensel JBM: Life-threatening bronchiolitis in children: Eight decades of critical care. Lancet Respir Med 2020; 8:142–144
3. Helmerhorst HJF, Schultz MJ, van der Voort PHJ, et al.: Bench-to-bedside review: The effects of hyperoxia during critical illness. Crit Care 2015; 19:284
4. Kallet RH, Matthay MA: Hyperoxic acute lung injury. Respir Care 2013; 58:123–141
5. Lilien TA, Groeneveld NS, van Etten-Jamaludin F, et al.: Association of arterial hyperoxia with outcomes in critically ill children: A systematic review and meta-analysis. JAMA Netw Open 2022; 5:e2142105
6. Rosenberg HF, Domachowske JB: Inflammatory responses to respiratory syncytial virus (RSV) infection and the development of immunomodulatory pharmacotherapeutics. Curr Med Chem 2012; 19:1424–1431
7. Kirolos A, Manti S, Blacow R, et al.; RESCEU Investigators: A systematic review of clinical practice guidelines for the diagnosis and management of bronchiolitis. J Infect Dis 2020; 222:S672–S679
8. Cunningham S, Rodriguez A, Adams T, et al.; Bronchiolitis of Infancy Discharge Study (BIDS) group: Oxygen saturation targets in infants with bronchiolitis (BIDS): A double-blind, randomised, equivalence trial. Lancet 2015; 386:1041–1048
9. Clark JM: Pulmonary limits of oxygen tolerance in man. Exp Lung Res 1988; 14:897–910
10. Balcarcel DR, Coates BM, Chong G, et al.: Excessive oxygen supplementation in the first day of mechanical ventilation is associated with multiple organ dysfunction and death in critically ill children. Pediatr Crit Care Med 2022; 23:89–98
11. Balasubramanian S, Suresh N, Ravichandran C, et al.: Reference values for oxygen saturation by pulse oximetry in healthy children at sea level in Chennai. Ann Trop Paediatr 2006; 26:95–99
12. Pediatric Acute Lung Injury Consensus Conference Group: Pediatric acute respiratory distress syndrome: Consensus recommendations from the Pediatric Acute Lung Injury Consensus Conference. Pediatr Crit Care Med 2015; 16:428–439
13. Bakdash JZ, Marusich LR: Repeated measures correlation. Front Psychol 2017; 8:456
14. Peters MJ, Jones GAL, Wiley D, et al.; Oxy-PICU Investigators for the Paediatric Intensive Care Society Study Group (PICS-SG): Conservative versus liberal oxygenation targets in critically ill children: The randomised multiple-centre pilot Oxy-PICU trial. Intensive Care Med 2018; 44:1240–1248
15. Schjørring OL, Klitgaard TL, Perner A, et al.; HOT-ICU Investigators: Lower or higher oxygenation targets for acute hypoxemic respiratory failure. N Engl J Med 2021; 384:1301–1311
16. Bem RA, Kneyber MC, van Woensel JB: Respiratory syncytial virus-induced paediatric ARDS: Why we should unpack the syndrome. Lancet Respir Med 2017; 5:9–10
17. Zhu Z, Zhou M, Wei Y, et al.: Time-varying intensity of oxygen exposure is associated with mortality in critically ill patients with mechanical ventilation. Crit Care 2022; 26:239
18. Suzuki S, Eastwood GM, Peck L, et al.: Current oxygen management in mechanically ventilated patients: A prospective observational cohort study. J Crit Care 2013; 28:647–654
19. Rachmale S, Li G, Wilson G, et al.: Practice of excessive F(IO(2)) and effect on pulmonary outcomes in mechanically ventilated patients with acute lung injury. Respir Care 2012; 57:1887–1893
20. Platen PV, Pomprapa A, Lachmann B, et al.: The dawn of physiological closed-loop ventilation-a review. Crit Care 2020; 24:121
21. Pannu SR, Exline M, Klamer B, et al.: Early titration of oxygen during mechanical ventilation reduces hyperoxemia in a pilot, feasibility, randomized control trial for automated titration of oxygen levels. Crit Care Explor 2022; 4:e0704
22. Schene KM, van den Berg E, Wösten-van Asperen RM, et al.: FiO2 predicts outcome in infants with respiratory syncytial virus-induced acute respiratory distress syndrome. Pediatr Pulmonol 2014; 49:1138–1144
23. Barbateskovic M, Schjørring OL, Krauss SR, et al.: Higher vs lower oxygenation strategies in acutely ill adults: A systematic review with meta-analysis and trial sequential analysis. Chest 2021; 159:154–173
24. Askie LM, Darlow BA, Finer N, et al.; Neonatal Oxygenation Prospective Meta-analysis (NeOProM) Collaboration: Association between oxygen saturation targeting and death or disability in extremely preterm infants in the neonatal oxygenation prospective meta-analysis collaboration. JAMA 2018; 319:2190–2201
25. Chang I, Thomas K, O’Neill Gutierrez L, et al.: Protocol for a randomized multiple center trial of conservative versus liberal oxygenation targets in critically ill children (Oxy-PICU): Oxygen in paediatric intensive care. Pediatr Crit Care Med 2022; 23:736–744

critical care; hyperoxia; mechanical ventilation; oxygen/blood; oxygen/therapeutic use; pediatric intensive care units

Supplemental Digital Content

Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.